Detection of Neuronal Activity in the Hippocampus of Sleep Deprived Rats Using Microelectrode Arrays

Zeying Lu ,Shengwei Xu ,Hao Wang ,Juntao Liu ,Fei Gao ,Yilin Song ,Jingyu Xie ,Guihua Xiao ,Yu Zhang ,Yuchuan Dai ,Yun Wang ,Lina Qu ,Xinxia Cai ,＊

1 State Key Laboratory of Transducer Technology,Institute of Electronics,Chinese Academy of Sciences,Beijing 100190,P.R.China.

2 University of Chinese Academy of Sciences,Beijing 100049,P.R.China.

3 State Key Laboratory of Space Medicine Fundamentals and Application,China Astronaut Research and Training Center,Beijing 100094,P.R.China.

Abstract:Sleep deprivation (SD)is the partial or complete loss of sleep and has long been used as a tool in sleep research to interfere with normal sleep cycles in rodents and humans.The progressivelyaccumulating sleep pressure induced by sleep deprivation can lead to a variety of physiological changes and even death.Compared to traditional detection methods,in vivo detection of neuronal activity using micro-electromechanical system (MEMS)technology following sleep deprivation can help fully elucidate the effects of sleep deprivation at the cellular level.Herein,a computer-controlled rotary roller was used to completely deprive rats of sleep for 14 days and 16-channel microelectrode arrays (MEAs)were fabricated and implanted into the rat hippocampus to measure neural spikes and local field potentials (LFPs)in real-time.The hippocampus is involved in learning and memory and has been the focus of intensive research aimed at understanding the function of sleep.This study was performed to measure the changes in neuronal activity in the rat hippocampus induced by sleep deprivation as well as their overall impact on the brain.After sleep deprivation,both the pyramidal- and interneurons showed a higher amplitude and more intense firing patterns.The fast-firing pattern of the neurons after sleep deprivation indicated elevated excitability in the prolonged awake state.In addition,the LFP of the sleep deprived rats fluctuated more frequently.The power of the LFPs in the low-frequency band (0-50 Hz)was calculated,showing increased power of the delta,theta,alpha,and beta bands after sleep deprivation,especially for the delta band (0.1-4 Hz).Generally,LFPs are generated by all types of neural activity in the neural circuit,and the changes in the low frequency band power suggested decreased arousal and increased sleep pressure induced by sleep deprivation,which could further impair brain function.This study was mainly aimed at measuring electrophysiological changes induced by sleep deprivation in the rat brain.Typically,neuronal activity changes were accompanied by the alternation of specific neurotransmitters in the brain.In the future,it will be essential to focus on measuring the concurrent change of electrophysiological and neurochemical signals to better examine the impact of sleep deprivation on brain function.

Key Words:Sleep deprivation; Neuronal activity; Microelectrode array; Spike; Local field potential

1 Introduction

Sleep is a complex physiological process and typically regulated by both cellular and molecular mechanisms in the organism1.Given its widespread presence and persistence among species in the animal kingdom,it is clear that sleep is vital for a lot of essential functions consisting of growth,energy storage,waste clearance,modulation of the immune system,information processing,memory,vigilance,and psychological condition2-5.

Sleep deprivation (SD)is a technique to realize the partial or complete loss of sleep in an organism and has been extensively used in sleep research to interrupt normal sleep circles in rodents and human to illuminate the function of sleep6.The progressively-accumulating sleep pressure brought by the sleep deprivation in the long term could induce a variety of physiological changes,such as hyperphagia,weight loss,hypothermia and impaired immune functions7-9.

The research about sleep deprivation was conducted by first maintaining the human or animal in the waking state,and then explaining the functions of sleep from these deprivation periods.Manaceine10showed that the death would occur in dogs following a few days of sleep deprivation and that the central nervous were severely damaged in these animals.Daddi11showed that the sleep deprivation could lead to neuronal degeneration in the spinal ganglia and cerebellar and cerebral cortices.Many researchers focused on the hippocampus in the brain to understand the sleep functions,since the hippocampus is a limbic structure involved in learning and memory12.Mohammed et al.6investigated the effect of loss of sleep on the amino acid neurotransmitters in the hippocampus of rats and found that neurochemical changes induced by sleep deprivation were associated with impaired brain functions.

The electroencephalogram (EEG)has been extensively applied in neurophysiological measures following total and partial sleep deprivation both in humans and animals13,14.The analysis of the EEG spectrum demonstrated that the sleep deprivation altered neuronal activities during sustained wakefulness15.Spectral power analysis has revealed increased power in theta,alpha and beta bands following sleep deprivation16.Recently,the delta activity has been the focus of the research about the effect of sleep deprivation on the EEG17.It has been suggested that the amount of time spent in prior wakefulness can often give rise to the increase in delta activity18,19.However,it is important to note that EEG recordings of the brain contain a wealth of information on neuronal activity,but are limited in that they only reflect gross neuronal firing patterns20,21.To emphasize the importance of understanding the underlying mechanism of sleep,we need to resort to the techniques that can more accurately monitor the neuronal firing activities at the cellular level.

As a high resolution and high sensitivity tool for neuronal research,the microelectrode arrays (MEAs)have been promoted by the recent advances in micro-electromechanical system(MEMS)technology22.In this study,we designed and fabricated a silicon-based 16-site implantable microelectrode array using MEMS techniques for in vivo measurement of neural discharge activity23.The neuroelectrical detections were performed in the hippocampus of normal and sleep deprived rats to study the impact of sleep deprivation on the brain24.Neural spikes and local field potentials (LFPs)in multichannels were acquired and analyzed25.The results showed the competence of the implantable MEA to record changes in neural electrical activity26and could verify the neurophysiological changes induced by sleep deprivation.

2 Experimental

2.1 Animals

In total n=8 adult male Sprague-Dawley rats weighing 220 to 250 g were used in this study (mean age 7 ± 0.5 weeks during the experiment).The rats were divided into two different groups,which are the SD group (n=4)and controlled group (n=4).Animals arrived at the laboratory one week before the onset of experiments to acclimatize to the environment and were housed individually in a light-controlled room (light on from 9:00 A.M.to 9:00 P.M.)with free access to food and water.Room temperature and relative humidity were maintained at (22 ± 1)°C and (50 ± 20)%,respectively.All animal procedures were performed with the permission and under the guidelines of the animal welfare committee.

2.2 Apparatus

A computer-controlled rotary roller was used to deprive the rats of sleep.A 128-channel neural data recording system(Blackrock Microsystems,USA)was used to detect the neural signal.Animal anesthesia during the experiments was performed by the animal isoflurane anesthesia machine (R580S,RWD Life Science,China).A micropositioner (Model 2662,David LOPF instrument,USA)was used to push the implantable MEA into the target area.Other apparatuses included a high-speed drill(carbon burr drill bits,0.7 mm,SAESHIN Ltd,Daegu,Korea)and a stereotaxic frame (Stenting,Wood Dale,IL,USA).

2.3 Design and modification of MEA

The recording electrodes used in our study were designed with four shanks,and each shanks contained four recording sites.The size of each shaft was 6 mm in length,100 μm in width and 200 μm in spacing.As shown in Fig.1a,four round sites (20 μm in diameter)were arranged as a diamond on the tip of each shaft,and 16 recording sites can record the electrophysiological activities simultaneously.The MEA was mass fabricated by MEMS thin-film technology in the super-clean laboratory as described in previous work24,29.

Fig.1 (a)Structure of the MEA,the 16-channel sites on the four shafts were used to detect electrophysiological signals;(b)the histological image with the trace of MEA implanted into the hippocampus.

The detection of neural signals is usually affected by the impedance of the electrode.The neural signals we detected are susceptible to the geometry and surface area of the exposed recording surface,as well as the materials used for recording.To reduce the electrode impedance and improve the signal-to-noise(SNR)ratio,all 16 recording sites were electroplated with Pt nanoparticles (PtNPs).This layer of PtNPs enlarged the recording surface area and decreased the impedance of the sites.PtNPs were electrodeposited onto the round recording site with chronoamperometry (CA,-1.1 V,30 s),in the plating solution involving 48 mmol·L-1chloroplatinic acid and 4.2 mmol·L-1lead acetate in a 1 :1 mixture30.

2.4 Design of experiment

The sleep deprivation procedures were carried out using a rotary roller.Experiments were conducted using the slow rotational movement of the rotary roller (programmed on a schedule of 1 min ‘on' at the speed of 60 s·r-1and 2 min ‘off')on the SD rats.As the sleep deprivation started,rats were put into the roller and maintained awake by being forced to move against the direction of rotation to avoid being falling down27.Similar parameters have been shown previously to keep the animals awake (98.9 ± 0.3)% of the time during a 24 h period of sleep deprivation on rats28.

Group 1 rats (n=4)served as SD group,and were first habituated to rotary roller initially and progressively for five days.For the first two days,mice were accustomed to the roller without motion.Mice were then in motion for 30 min per day at 10:00 in the next three days.The day after wheel habituation sessions,the rats were exposed to the sleep deprivation for 14 days.

Group 2 rats (n=4)were set as the control group and were not subjected to sleep deprivation.Before the experiment started,animals in the control group were put into the SD apparatus and subject to the habituation procedure as group 1.For the next 14 days,rats were housed in the SD apparatus and conducted the same physical activity as the SD group except that they had access to sleep from 06:00 to 18:00 when no rotary movement was on.

2.5 Surgical procedures

All rats were anesthetized by R580S isoflurane anesthesia and fixed on the stereotaxic instrument.The craniotomies were conducted to conform to the recording site (AP(anteroposterior):-3.0 mm,ML (mediolateral):2.0 mm)for the implantation of the MEA.One more site was drilled to establish a ground connection.As shown in Fig.1b,the histological image showed the MEA was vertically implanted into the hippocampus of the rat.The 16 recording sites on the four shafts could simultaneously monitor the neural spikes and LFPs changes in the hippocampus region.

2.6 Data acquisition and analysis

All the recording processes were performed under the anesthesia of the rats.For the prevention of external electromagnetic interference,the stereotaxic frame with the rat was put into a grounded,shielded box during the measurements.The implanted MEA was connected to a 128-channel electrophysiological data acquisition system to detect the neural activity.The sampling frequency was set as 30 kHz to continuously acquire the raw data.Both the LFP data (low pass filter of 100 Hz)and neural spike recordings (high pass filter of 500 Hz)were extracted from the neurophysiological data on the MEA simultaneously.After the measurements,the NeuroExplorer and Offline Sorter softwares were used for the analysis of electrophysiological data29,31.

3 Results and discussion

3.1 Neurophysiological recording

Before the experiment,a certain threshold would be set in order to detect the waveforms of unique shapes from the signal above the background noise.The waveforms we detected should be easily recognized with the naked eye without any further signal processing.Usually,Signal-to-noise ratio (SNR)offered a reliable way to perform quantitative measures.Generally,the signals we detected in this measure have SNR above 5.

In the experiment,waveforms of about 10 channels have sufficient amplitude to be triggered above background noise.Typically,the neurons in the rat hippocampus could be classified into different types,which mainly include the pyramidal neurons and interneurons and act differently.Fig.2 and 3 showed the neural spike trains of Pyramidal neurons and Interneurons and simultaneously recorded LFP of six different channels under anesthesia in the time segments of 30 s.Compared with the normal rats,the spike of both the Pyramidal neurons and Interneurons after sleep deprivation showed a more intense firing Pattern and the amplitude of spike was higher as well.Meanwhile,the LFP of the SD rats fluctuated more frequently,indicating that the sleep deprivation exerted excitatory effects on the neurons in the hippocampus.

3.2 Analysis of neural spike

Fig.2 (a)The neural spike firing trains of Pyramidal neurons and Interneurons recorded in the hippocampus of normal rats and(b)SD rats under anesthesia in the time segments of 30 s.

Fig.3 Six channels (Ch3,Ch5,Ch10,Ch11,Ch14,Ch15)of(a)LFPs recorded in the hippocampus of normal rats and(b)LFPs recorded in the hippocampus of SD rats under anesthesia.

Fig.4a showed the average spike waveforms detected in the hippocampus of the normal and SD rats.The spikes of the normal rats had a peak latency of (0.69 ± 0.06)ms.For the SD rats,the peak latency of spikes decreased to (0.41 ± 0.04)ms.The reduction of the peak latency suggested a more swift alternation of the spike and might be related to the high-firing pattern of the neurons in SD rats.Fig.4b showed the normal distribution curves of the firing rate for the neurons of the normal and SD rats in the frequency domain.We found that the neurons of the normal rats were mainly firing in the low frequency delta band (0.1-4 Hz),while the spike firing of the SD rats were mainly concentrated in the theta band (4-8 Hz).

We further calculated the amplitudes and firing rates of the pyramidal neurons (n=10)and interneurons (n=14)we detected respectively.Fig.4c showed the amplitude of the pyramidal neurons and interneurons.For the normal rats,the average spike amplitude of the pyramidal neurons was 113.2 μV.We found that the amplitude increased by 15.0% to 130.2 μV after sleep deprivation.And the spike amplitude of the interneurons increased by 23.3% from 123.1 to 151.6 μV after sleep deprivation.As shown in Fig.4d,the average firing rate was also calculated.For the normal rats,the firing rates of the pyramidal neurons and interneurons were 0.8 and 1.8 Hz,respectively.After sleep deprivation,the firing rates of the pyramidal neurons and interneurons increased by 102.4% and 46.3% to 1.7 and 2.6 Hz,respectively.

3.3 Power spectral density analysis of the local field potentials

The local field potential (LFP)was defined as the electrophysiological signal generated by the summed electric current flow due to the synaptic activity of neurons within a small volume of tissue.The voltage is produced across the local extracellular space by action potentials and graded potentials in neurons in the area.For this reason,through the computation of the power spectrum,we can better quantify the LFP activity.The power spectrum is commonly defined as the Fourier transform of the autocorrelation function and describes the distribution of power into frequency components composing that signal.Generically,the LFP was dominated by the low-frequency band.So we mainly analyzed the frequency band under 50 Hz of LFPs recorded from the normal and SD rats under anesthesia.In Fig.5a,we put the power spectrum density (PSD)in the hippocampus of normal and SD rats in one coordinate system and analyzed their relative features.We found that there was an overall increase in the power spectral density of frequency band under 50 Hz after sleep deprivation compared to the normal.The power of LFPs was subsequently calculated and shown in Fig.5b.We found there was an increase from (4.25 ± 0.09)mW to (10.22 ± 0.15)mW after sleep deprivation,which was twice the power of LFPs in the normal rats.Then the power of each frequency band was calculated and the results were shown in Table 1.It revealed an increase in the power of delta,theta,alpha and beta bands after sleep deprivation,especially in the power of the slow frequency delta band,which was more than three times than that of the normal rats.

Fig.4 (a)The average spike waveforms sorted from 4236 spikes in the normal rats and 3648 spikes in the SD rats; (b)the normal distribution curves of firing rate for the normal and SD rats in the frequency domain; (c)the average spike amplitude (peak to peak)of the pyramidal neurons and interneurons in normal and SD rats; (d)the average spike firing rates of the pyramidal neurons and interneurons in normal and SD rats.

Fig.5 (a)The PSD of LFPs of normal and SD rats in the frequency band of 0-50 Hz,the shadow is the error bar calculated from the PSD of the LFPs recorded by the six channels; (b)the average power of LFPs of normal and SD rats in the frequency band of 0-50 Hz.the shadow is the error bar calculated from the PSD of the LFPs recorded by the four channels.(＊＊P ＜ 0.01).

Table 1 The relative frequency band power of LFPs in normal and SD rats.

It is well known that prolonged waking enhances the propensity of sleep,while extended sleep raises the likelihood of waking.The higher homeostatic pressure is associated with higher slow wave density,higher amplitudes,and steeper slopes of the slow oscillations.In this study,the increase in delta activity after sleep deprivation may reflect the accumulation of sleep pressure in rats32,33.Besides,several studies demonstrated the important role of brain activities of theta wave range in information processing34.Increased theta activity is found to correlate with the increased workload,as well as increased fatigue35.The augmentation of theta activity after sleep deprivation might be associated with diminished cognitive ability,which could induce learning-deficit behaviors of the rats.

Importantly,we found that sleep deprivation activated a substantial proportion of hippocampal Pyramidal neurons and interneurons,which could lead to an overall increase in neuronal activities in this area.The mechanisms underlying the fast-firing patterns of the neurons in the hippocampus after sleep deprivation may be complex.As is well known,movement and behavioral state transitions could indicate the widespread changes in the activity of several neuromodulatory systems36.In the previous study,increased motor activity following the sleep deprivation could be induced in the animal model37,as well as depressed human38.The behavioral experiments showed that the rats which performed post sleep deprived training exhibited impaired hippocampus-related brain functions,such as learning,memory and spatial navigation39.It was possible that the constant waking state increases the excitability of neurons in the hippocampus,and thus the decreased arousal and increased sleep pressure induced by prolonged waking behavior could further impair the brain function,which may contribute to the deficits in learning and memory behavior after sleep deprivation40.

On the other hand,in vivo studies suggested that the changes of specific amino acid neurotransmitters in the hippocampus and cortex after sleep deprivation may modulate the neuronal excitability.In a previous pilot study,experimental rats deprived of more than five days could lead to their death,and the changes in both GABA and glutamate levels became dominant after the first 24 h of sleep deprivation.These changes were sustained during the five days of sleep deprivation until the death41.Bettendorf et al.also reported that elevated content of glutamate and glutamine was found in the rat cortex after sleep deprivation42.Glutamate and gamma-aminobutyric acid (GABA)are neurotransmitters involved in the nervous system regulation,which could exert excitatory and inhibitory effects on the neurons,respectively.The findings suggested that increased levels of glutamate may overweigh the effects of GABA on the neurons during sleep deprivation41.It is thus assumed that the prolonged waking behavior could elevate the excitatory neurotransmitters in the brain levels,which may keep the neurons in the constant activating state.

4 Conclusions

The objective of this work was to observe and try to elucidate the electrophysiological changes following 14-day total sleep deprivation.The 16-channel MEA was used to detect the changes in neuronal activities after the sleep deprivation.According to our analysis,the amplitude and firing rates of both Pyramidal neuron and Interneuron spike increased after the sleep deprivation.Besides,there was an overall increase in the power of the slow frequency band (0-50 Hz)of the PSD data in the SD rats compared to the normal rats in the present study.It revealed that the sleep deprivation increased the excitability of the neurons and lead to decreased arousal and increased sleep pressure,which could further impair the brain functions.Usually,the changes in neuronal activities were accompanied by the alternation of specific neurotransmitters in the brain.In our future study,more hippocampus-related behavioral tasks would be further performed to assess the functional loss caused by sleep deprivation and the concurrent change of electrophysiological and neurochemical signals would be measured to better illustrate the impact of sleep deprivation on the brain functions at the cellular level.